system
by
Elizabeth Johanna Opperman
Thesis presented in partial fulfilment of the requirements for the degree of
Master of Entomology
at
Stellenbosch University
Department of Conservation Ecology and Entomology, Faculty of AgriSciences
Supervisor: Professor John Terblanche Co-supervisor: Dr Minette Karsten
March 2018
The financial assistance of the National Research Foundation (NRF) towards this research is hereby acknowledged. Opinions expressed, and conclusions arrived at, are those of the author
Declaration
By submitting this thesis electronically, I declare that the entirety of the work contained therein is my own, original work, that I am the sole author thereof (save to the extent explicitly otherwise stated), that reproduction and publication thereof by Stellenbosch University will not infringe any third party rights and that I have not previously in its entirety or in part submitted it for obtaining any qualification.
Date: 26/01/2018
Copyright © 2018 Stellenbosch University All rights reserved
3
Summary
The mechanisms and traits influencing insect invasions are generally poorly understood. Drosophilids are an excellent model system for studying invasions and especially the adaptive processes occurring during invasions since the family has short generation times, diverse functional traits and variation in geographic distributions while possessing several notable invasive species. While there are many studies of environmental stress resistance or life-history traits and how these might influence population dynamics or geographic range limits in
Drosophilidae, these studies have several potential shortcomings. Chief among these are perhaps concerns surrounding their use of Stock Centers (laboratory cultures) of varying time in culture and sometimes unknown geographic origins to infer trait-environment associations, niche requirements or evolutionary adaptive capacity. Traits can respond rapidly to laboratory rearing with laboratory cultures typically losing stress resistance and increasing fecundity and/or development rates. In this study, I sought to determine whether there is a significant and
systematic effect of time spent in culture on estimates of environmental stress resistance and its thermal acclimation (i.e. phenotypic plasticity) of two wild-caught Drosophila species
(Drosophila melanogaster and Zaprionus vittiger) between newly established lines (in the F2
generation) and a later timepoint (F8-F10 generation) in the laboratory under standard, controlled
rearing conditions. A further objective was to identify the nature and magnitude of basal and plastic estimates of environmental stress resistance traits among four populations of D.
melanogaster collected from different areas within South Africa to assess if geographic origin
influences trait and plasticity estimates substantially within a single species. This was done by measuring traits of upper and lower thermal activity limits (CTMAX and CTMIN, respectively), the
proportion of individuals surviving after 24 hours after exposure to a potentially lethal
temperature (heat and cold survival survival), desiccation resistance, starvation resistance and the plasticity thereof in response to thermal acclimation at three temperatures (18, 23, 28 ˚C). There was significant variation in resistance to environmental stressors between earlier and later generations for D. melanogaster and Z. vittiger. Drosophila melanogaster generally increased resistance to environmental stressors after spending ten generations in the laboratory whilst Z.
4
survival traits and the plasticity thereof between the four populations of D. melanogaster. Thus, it is clear that conditions at time of sampling and the species or population’s geographic source can strongly mediate trait and plasticity assessments in laboratory cultures. Consequently, environmental stress resistance measured from Stock Centers lines or species may give a biased view which could influence tests of climate or niche matching and risk assessments. The
divergent, idiosyncratic responses noted between my study species’ means that more species would need to be assessed to understand the generality of the outcomes described here.
5
Opsomming
Meganismes en eienskappe wat suksesvolle indringings te weeg bring word nie goed verstaan nie. Die Drosofiliede verteenwoordig ‘n belangrike model sisteem vir die studie van
indringerspesis en die adaptiewe prosesse wat gepaard gaan met suksesvolle indringings, omdat die familie diverse funksionele eienskappe besit, ‘n wye geografiese verspreding het en
terselfdetyd uit verskeie kenmerkende indringerspesies bestaan. Daar bestaan verskeie studies wat die omgewings stres weerstands eienskappe van die Drosofiliede gemeet het, maar hierdie studies het verskeie tekortkominge. Een van die belangrikste tekortkominge is moontlik bekommernisse rondom die gebruik van ‘Stock Centers’ (laboratorium kulture) van verkeie ouderdomme en geografiese afkoms om omgewings assosiasies te maak of om evolutionêre adaptiewe kapasiteit te bepaal. Spesies eienskappe reageer vinnig op laboratorium kondisies en verloor tipies hulle weerstand tot omgewings kondisies met ‘n gepaardgaande toename in voortplanting. Die doel van hierdie studie was om te bepaal of die aantal tyd wat spandeer word in kultuur ‘n beduidende effek het op beramings van omgewings stress weerstand en termiese akklimasie (fenotiepiese plastisiteit) tussen nuutgestigte (F2 generasie) Drosofilied spesies
(Drosophila melanogaster en Zaprionus vittiger) en na die spesies tien generasies (F10 generasie)
in kultuur onder standaard grootmaak praktyke spandeer het. ‘n Verdere doel van die studie was om die natuur en omvang van basale en plastiese beramings van omgewings stress weerstand eienskappe te identifiseer binne vier populasies van D. melanogaster wat gevang was in verskeie dele van Suid Afrika om sodoende te bepaal of geografiese afkoms ‘n beduidende impak kan hê op beramings van omgewingseinskappe en plastisiteit binne ‘n spesifieke spesies. Hierdie was gedoen deur die boonste en onderste termiese limiete (CTMAX en CTMIN), die proporsie van
individue wat oorleef het 24 uur na blootstelling aan ‘n potensiële dodelike temperatuur (hitte en koue skok) sowel as oorlewing na uitdroging en uithongering, en hulle plastisiteit in reaksie tot termiese akklimasie by drie temperature (18°C, 23°C en 28°C), te bepaal. Daar was
betekenisvolle en teenstrydige weerstand tot omgewing stressors tussen vroeër en latere generasies van D. melanogaster en Z. vittiger. Drosophila melanogaster het ‘n algemene toename in weerstand tot omgewing stressors gehad na tien generasies in die laboratorium, terwyl Z. vittiger ‘n afname ondergaan het. Daar was ook betekenisvolle verskille in beide termiese en oorlewingseienskappe sowel as hulle plastiese reaksie tussen die vier populasies van
6
D. melanogaster, en dus is dit duidelik dat die geografiese oorsprong van ‘n spesies of populasie
‘n effek kan hê op die assessering van eienskappe en hulle plastisiteit in laboratorium kulture. Dus is dit duidelik dat die kondisies tydens steekproefneming sowel as die spesies of populasie se geografiese oorsprong ‘n verdere invloed het op omgewingseienskap en platisiteit assesserings in laboratorium culture. As ‘n gevolg sal omgewings stress eienskappe wat gemeet is vanaf ‘Stock Centers’ ‘n bevoordeelde uitkyk gee en hierdie vooroordele kan ondersoeke van klimaat en nis ooreenstemmings beïnvloed en sodoende risiko assessering beïnvloed. Die uiteenlopende eienaardige reaksies opgemerk tussen my studie spesies beteken dat meer spesies geasseseer sal moet word om die algemeenheid van die uitkomste wat hier beskryf word te verstaan.
7
This thesis is dedicated to
8
Acknowledgements
I wish to express my sincere gratitude and appreciation to the following persons and institutions: My supervisors, Prof. John Terblanche and Dr Minette Karsten without whom this work would not have been possible. I thank them for their guidance, advice, criticism, and encouragement as well as multiple reviews of this thesis.
I would also like to thank Dr Stefan Foord and Remember Baloyi from the University of Venda as well as Professor Des Conlong and his team from SASRI, Saskia Thomas, Minette Karsten, Mr and Mrs Agenbag and Liana De Araujo for helping with fly collections. Additional thanks to Izak and Annelene Huisamen and Frikkie and Jeanette van As for allowing me to catch flies at their homes and for their enthusiastic help in trying to find different species. It is greatly appreciated.
Dan wil ek ook dankie se aan my ma (Elsabè Opperman), broer (Christiaan Opperman) en ouma (Bettie Bothma) vir hulle hulp en ondersteuning tydens my projek. Laastens wil ek baie dankie sê aan Dawid Huisamen vir sy hulp met vlieë vang, vir ontelbare kilometers se rondry, vir saam uitkamp om vlieë te vang en vir die uithou met my hoë stresvlakke. Ek waardeer al die ekstra werk, vroeg opstaan en min slaap opreg.
Ek wil ook baie dankie sê aan God wat my leiding gee en dra en my in staat stel om te werk tot ere van Sy naam.
9
Preface
This thesis is presented as one chapter
Contents
Introduction ... 11
Invasive species ... 11
Drosophilidae as invasive species ... 12
Estimating environmental tolerances and inferring population dynamics ... 14
Variation in trait estimates ... 16
Stock Centers and laboratory adaptation ... 18
Study objectives ... 19
Aims: ... 20
Materials and Methods ... 21
Origin and maintenance of experimental flies ... 21
Environmental stress resistance ... 23
Temperature traits ... 23
Heat and Chill survival ... 24
Survival traits ... 24 Statistical analyses ... 25 Results ... 26 Temperature treatments ... 26 Drosophila melanogaster ... 26 Zaprionus vittiger ... 27
Heat and Chill survival ... 30
Drosophila melanogaster ... 30 Zaprionus vittiger ... 31 Survival treatments ... 34 Drosophila melanogaster ... 34 Zaprionus vittiger ... 35 Populations of D. melanogaster ... 40
10
Temperature treatments... 40
Heat and Chill survival ... 43
Survival treatments ... 45
Discussion ... 49
References ... 60
11
Introduction
Invasive species
Species introduced and established intentionally or unintentionally beyond their native range are referred to as alien invasive species (Jeschke et al., 2014). Alien invasive species can have numerous detrimental impacts on the natural world, and act as agents of ecological change by causing changes in ecosystem structures and the extinction of threatened species, by altering the structure of communities and by disrupting successional pathways (Clout and Williams, 2009; Jeschke et al., 2014; Simberloff et al., 2013). Invasive species can enter a country through several pathways, either intentionally or unintentionally, and in recent decades there has been an increase in the spread of invasive species through human vectors correlated with an increase in human movement across the world (Roderick and Navajas, 2015; Sacaggi et al., 2016). Invasive species also cause losses of important ecosystem services with accompanying impacts on the health of humans, livestock and wild animals (Bertelsmeier et al., 2016; Clout and Williams, 2009). For a species to become invasive, it needs to overcome several barriers to survive through the invasive stages of transport and introduction, and once introduced, population establishment and eventual spread (Blackburn et al., 2011). A suite of population demographic factors and traits likely influence the invasion process. Upon introduction to a new environment, high propagule pressure (the number of individuals introduced and frequency thereof) increases the chances of successful invasion (Duncan et al., 2014). However, both demographic and genetic characteristics of the introduced species are chief in allowing spread of the species after the initial introduction (Szűcs et al., 2014). A match between the introduced areas’ climate and the environmental stress traits of the introduced species allows for niche occupation and increased population growth and spread of the invasive species post-introduction (Dixon et al., 2009; Gilchrist et al., 2008; Rey et al., 2012). Thus, environmental stress resistance or thermal requirements of a species represent an important proxy for determining which species have the potential to become invasive and are frequently used in risk assessments and climate or niche models (e.g. Jarošík et al., 2015; Kumschick et al., 2015).
Invasive insects have detrimental ecological and social impacts worldwide, however, Pysêk et al. (2008) found that studies on insects make up only 18% of invasive studies. Relative to insect
12
species diversity (c. 5.5 million species) this is a potentially huge literature bias. Well-known insect invaders include the Argentine ant (Linepithema humile), the big-headed ant (Pheidole
megacephala), the gypsy moth (Lymantria dispar), the codling moth (Cydia pomonella), the
Mediterranean fruit fly (Ceratitis capitata) and Drosophila suzukii among others (Huang et al. 2011). The Argentine ant, for example, has spread globally through the increased movement of people and trade goods (Suarez et al., 2001) and has reduced native ant populations by causing the collapse of various mutualisms between plants and native ant communities (Griffiths and Picker, 2011). In Fynbos vegetation in South Africa, the Argentine ants interfere with the burial of large proteaceous seeds by native ants, leading to changes in the plant community structure (Griffiths and Picker, 2011; Mothapo and Wossler, 2013).
In addition, the true fruit flies (Tephritidae), such as members of Ceratitis and Bactrocera, have led to substantial economic losses in the fruit industry as well as quarantine restrictions and phytosanitary measures in affected areas that lead to increased costs of fruit production and restricted market access (Sarwar, 2015). The Mediterranean fruit fly (Tephritidae) represents a major fruit fly pest which has readily invaded most of the world (Malacrida et al., 2007) and which has become an important agricultural pest with associated economic losses for the fruit industry worldwide (De Meyer et al., 2008; White and Elson-Harris, 1994). Their fast spread is due to the increased implementation of agriculture as well as the increased mobility of humans worldwide (Hill et al., 2016; Malacrida et al., 2007).
Drosophilidae as invasive species
Several invaders from the Drosophila genus have also become important pests in the fruit industry. The most important include Drosophila subobscura, Zaprionus indianus and D.suzukii (the spotted wing Drosophila). Drosophila subobscura (native to Europe) has invaded both South and North America leading to severe economic and ecological impacts (Foucaud et al., 2016). Zaprionus indianus (native to East Africa, the Middle East and Southern Eurasia) has invaded India, Mexico and Canada among others and has also had detrimental impacts on the fruit industry in these countries (Alawamleh et al., 2016; Van der Linde et al., 2006). Drosophila
already-13
damaged fruit (Lasa and Tadeo, 2015). Drosophila suzukii is originally from Japan but has subsequently invaded China, Myanmar, India, Italy, Thailand, Spain, Russia and North America where it causes severe damage in the fruit industry (Calabria et al., 2012; Walsh et al., 2011). It was first identified in 2008 from raspberries in California (USA) but an accurate identification was only made in 2009 after it had spread to many countries worldwide. It is unique in that it is a primary pest, unlike other Drosophila species, with a serrated ovipositor that allows it to lay its eggs in undamaged fruit and, consequently, leads to substantial fruit losses in the invaded areas (Calabria et al., 2012). It is a major pest on cherries, strawberries, pears and wine grapes, among others (Calabria et al., 2012; Walsh et al., 2011).
In order to become invasive, species are thought to go through four stages, namely transport, introduction, establishment and spread (Blackburn et al., 2011). The species will only become invasive if it survives all these stages overcoming the unique barriers posed in each instance (Blackburn et al., 2011). Successful invasions require an organism to have the ability to respond to diverse environmental conditions through either phenotypic plasticity or rapid genetic
adaptive shifts (Perkins et al., 2011). As a result, Drosophila species act as an important model for investigating the underlying mechanisms of adaptive processes involved in successful
invasions (Gibert et al., 2016) due to their wide geographical distribution as well as the presence of latitudinal clines for many morphological and physiological traits. In addition, they are easy to rear, have short generation times (leading to rapid evolutionary shifts) and there is a large
amount of genomic data available for the genus (Gibert et al., 2016; Hoffmann, 2010).
Risk assessments determine the likelihood of a species becoming invasive together with the potential impact should it become invasive (Kumschick and Richardson, 2013). Except for risk assessments done on the widespread pest species D. suzukii, little risk assessments have been done on other potential Drosophilidae invaders (Berry, 2012). This is due to a lack of
prioritization of members of the family, because of limited data availability and that the
particular species’ invasion may have little economic or ecological consequence (Kumschick et
al., 2015).Data required for risk assessments typically includes propagule pressure as well as inherent biological variables (suitability of the climate, availability of resources and habitat, etc.) among others (Kumschick et al., 2015; see also Jarošík et al., 2015). These data can be used to
14
conduct a risk assessment for Drosophila species using a scoring system, like the newly developed generic impact scoring system (GISS) that will aid in elucidating potential future invaders (Nentwig et al., 2016).
Estimating environmental tolerances and inferring population dynamics
Several theories have been put forward on what characteristics or traits makes a species a successful invader. These include propagule pressure, genetic similarity (species from the same families and/or genus), the ability of the species to disperse long distances and climate matching among others (Richardson and Pysêk, 2006). In weeds for example, traits of high reproductive potential such as autonomous seed production (uniparental reproduction) and higher germination rates increases invasive potential (Hao et al., 2011; van Kleunen et al., 2007), together with performance traits such as higher shoot- and leaf area allocation and increased growth rate (van Kleunen et al., 2010; van Kleunen et al., 2011) all increase invasive potential. Several traits make invertebrates successful invaders, including propagule pressure (Hee et al., 1999), traits of high reproductive potential such as a high number of offspring and a wide host range (Duncan et
al., 2014) and long distance anthropogenic dispersal as a result of trade (fruit, flowers and the
like) (Brown et al., 2011) among others. Climate matching puts forward the idea that the climate in an organism’s native range can be compared to climates across the world in order to determine possible areas in which the organism could become invasive (Baker et al., 2000; Bomford et al., 2009; Petersen, 2003; Thorn et al., 2009). As a consequence, information on various aspects of the environmental physiology of Drosophila species are critical to ensure accurate risk
assessment of potential invaders which will also allow for more accurate climate modelling to determine important risk areas for invasion.
Critical thermal limits are thought to represent the functional, though not necessarily lethal, limits to performance of insects. Estimates of critical thermal limits are frequently estimated and used in database compilations of upper thermal tolerance or geographic distribution (reviewed in Terblanche et al., 2011) and are defined as the temperature at which the insect’s movements becomes irregular or loses function (e.g. righting response), causing it to be unable to escape the conditions that will ultimately lead to its death (Lutterschmidt and Hutchison, 1997).
15
Internationally, many physiological studies have used Drosophila as a model group to understand their trait-environment relationships, geographic range limits, range shifts or evolutionary adaptations (Bush et al., 2016; Hoffmann et al., 2002; van Heerwaarden et al., 2009; van Heerwaarden et al., 2016). Recent examples include Kellerman et al. (2012a) which assessed patterns of cold tolerance and desiccation resistance of 95 Drosophila species and found that desiccation and cold resistance are indeed linked to species distributions. Using the same species, they then compiled critical thermal maximum (CTMAX) and critical thermal minimum
(CTMIN)measurements (Kellerman et al., 2012b) and found variation in upper thermal limits
between the species and showed that species from drier regions had increased resistance to heat stress. In terms of thermal traits for Drosophila, several mechanisms have been studied including high and low temperature pre-treatments (acclimation), that lower thermal limits are generally more plastic than upper thermal limits (Kellett et al., 2005), a linear reaction norm for both CTMAX and CTMIN across acclimation temperatures (Schou et al., 2017) and significant
differences in basal thermotolerance and their plastic responses (Nyamukondiwa et al., 2011). Other studies assessing CTMAX and CTMIN and their environmental trait interactions found clinal
patterns in D. melanogaster thermal traits in Australia and showed that simulating temperate conditions in five widespread and five restricted Drosophila species increased CTMIN by 2-4°C
whilst simulating tropical conditions increased CTMAX by less than 1°C (Hoffmann et al., 2005b;
Overgaard et al., 2011). Furthermore, Matzkin et al. (2011) showed that phylogenetic relatedness has a large impact on both desiccation and starvation resistance. Drosophila melanogaster was also found to exhibit clinal patterns in desiccation resistance as well as cuticular permeability and levels of melanisation (darker versus lighter flies) (Bazinet et al., 2010; Hoffmann et al., 2003; Parkash et al., 2008).
While there are many physiological studies available, little is known about Drosophila species and their stress resistance in South Africa. Nyamukondiwa and Terblanche (2010) collected
Zaprionus vittiger individuals from Stellenbosch and showed that maximum survival was
achieved following cold-hardening at 7°C and 10°C. Klepsatel et al. (2013) investigated the thermal reaction norms between different populations of D. melanogaster sampled from South Africa, Ethiopia, Zambia, Switzerland and Austria and found no significant differences between the thermal reaction norms of the different populations for the morphological and reproductive
16
traits (fecundity, thorax length, wing area and ovariole number) they examined. To my knowledge, no other physiological studies have been done on Drosophilid species caught in South Africa. Other studies of African Drosophila species are mostly genetic or for compilation in large-scale population genetics or genomics. For example, Singh et al. (1982) investigated the genetic differences between D. melanogaster populations from nine continents which included a West African (Benin) population. They found significant genetic differentiation between the different populations, especially between the Northern and Southern hemispheres. Capy et al.
(1993) examined morphometric differences between D. melanogaster and Drosophila simulans populations and included populations from South Africa (Johannesburg and Cape Town) as well as Egypt. This study showed morphometric differences between populations of both D.
melanogaster and D. simulans, as well as between the Northern and Southern hemispheres for
both species.
Variation in trait estimates
There is a long history of comparative studies of environmental stress responses within and among Drosophilid species (e.g. Castaneda et al.,2015; Hoffmann and Harshman, 1999;
Hoffmann and Parsons, 1993; King et al., 1956; Lockwood et al., 2017; Smith and Smith, 1954). In nature, variation among seasons at specific geographic locations can elicit significant trait variation. Resistance to environmental stressors has been found to differ seasonally as well as geographically between populations of the same species (Hoffmann and Harshman, 1999; Sinclair et al., 2012). In Australia, desiccation resistance of D. melanogaster was found to be higher under summer versus winter conditions (Hoffmann et al., 2005b). Similarly, in India, seasonal changes in moisture availability led to changes in desiccation resistance as well as body colour in Drosophila jambulina (Parkash et al., 2009). Moreover, Behrman et al. (2015) found that D. melanogaster isofemale lines started from the generation that survived through the winter (females collected in Spring) had a higher tolerance to environmental stressors in comparison to isofemale lines started from females collected in the other seasons.
Since there is marked genetic variation in several traits from the cosmopolitan D. melanogaster, differences in environmental stress responses are largely expected between different populations
17
(Ayrinhac et al., 2004; David and Capy, 1988). There are several studies showing geographic variation in trait estimates between different populations of diverse Drosophilid species (Sinclair
et al., 2012). Recently, Sgrò et al. (2010) showed latitudinal clines in heat tolerance among
populations of D. melanogaster with populations at the tropics being more resistant to heat than those at higher latitudes. Earlier clinal studies documented increased heat resistance and
decreased cold resistance in lower latitude populations compared to higher latitude populations (Hoffmann et al., 2002; Hoffmann et al., 2005b). In Drosophila buzzatii clinal variation was found in thermal traits between populations along an altitudinal gradient in North-Western Argentina (Sørensen et al., 2005). Knockdown time following a heat shock treatment is higher in Australian D. melanogaster populations closer to the tropics in contrast with populations in temperate habitats. In addition, populations from warm regions are more heat tolerant and
populations from cold areas are more cold tolerant in D. melanogaster from temperate (Denmark and Italy) and subtropical areas (Canary Islands and Mali) (Guerra et al., 1997; Hoffmann et al., 2002). With regards to other physiological traits of stress resistance, significant differences in desiccation and starvation resistance have been widely reported between populations of the same species in D. melanogaster, D. ananassae and Zaprionus indianus in India (Karan and Parkash, 1998), D. birchii in Australia (Hoffmann et al., 2003) and in D. melanogaster, D.
pseudoobscura, D. nigrospiracula and D. mojavensis from several countries (Matzkin et al.,
2007).
Inter-specific variation in environmental stress resistance traits are also well-documented. For thermal traits, precipitation and temperature are thought to influence CTMAX values in several Drosophila species (Kellermann et al., 2012b). A study by Anderson et al. (2014) also found that
for 14 Drosophila species, the thermal traits of CTMIN, lower lethal temperature and lethal time at
low temperature were the best predictors of latitudinal distributions in Drosophila species worldwide. Additionally, widespread species were found to have higher cold resistance in Australian Drosophilacompared to narrowly distributed Drosophila species (Overgaard et al., 2011). Substantial variation in the heat acclimation response of several Drosophila species has also been found (Schou et al., 2017). Differences between Drosophila species are also
documented for desiccation and starvation resistance (Matzkin et al., 2009) with desiccation resistance correlated with distribution in some Drosophila species (Kellermann et al. 2012a).
18
These studies indicate the importance of taking both seasonal and geographic differences
between populations of the same species into account when assessing stress resistance traits for a species.
Stock Centers and laboratory adaptation
Although physiological studies on Drosophilidae are abundant there are several potential shortcomings in these studies (Hodkinson, 2003) that are typically ignored or argued to be of little consequence to the major outcomes or conclusions reached (Chown et al., 2003). For example, many inter-specific comparison studies make use of Stock Center lines, instead of newly established lines of flies. This is much less of an issue in inter-population studies which more typically would establish new lines from field collections (Hoffmann et al., 2001a; Kellermann et al., 2017; Sgró et al., 2010). Using species obtained from Stock Centers for experiments, instead of species collected from the wild, has the drawback of the laboratory colony potentially being inbred with a resultant decrease in genetic diversity (Ærsgaard et al., 2015) and possibly trait diversity. Additionally, the species could have become laboratory adapted potentially leading to a decrease in resistance to environmental stressors. Studies investigating these effects show that in laboratory stocks resistance to environmental conditions is lost and is instead replaced by selection on other traits such as increased fertility (Hoffmann et
al., 2001b; Sgró and Partridge, 2000). Species kept in the laboratory for an extensive period are
not subjected to natural selection and thus the results of stress resistance might not be a true representation of what may be happening in the wild if strong directional selection maintains the trait. Yet different species of Drosophila have vastly different responses to selection depending on the underlying adaptive capacity of the trait in question (Kellermann et al., 2009; van Heerwaarden and Sgrò, 2014), thus it is unlikely that all the species and diverse traits used in comparisons from Stock Center lines have responded equally to these laboratory rearing environments. To improve the validity of data collected for risk assessment, wild collected species, kept under laboratory conditions for a short period of time would better represent the wild population as they will more closely match the wild populations (Najarro et al., 2015).
19
Study objectives
This study aimed to investigate whether there is significant variation in environmental stress resistance traits in newly established laboratory lines (i.e. at the 2nd generation) compared to lines that have been in laboratory culture for a longer period (i.e. several generations). I chose the F2
generation to represent the wild population and F8-F10 to represent a time-point after laboratory
stay. I assumed that the results between F8-F10 would not significantly differ but that most
variation that would be a consequence of lab culture or lab adaptation would likely occur within the first 3-6 generations under standard rearing conditions (see e.g. Bertoli et al., 2009;
Sambucetti et al., 2010). To achieve this, I compared the same sets of traits scored recently after establishment and again later on in two wild-caught South African Drosophila species (D.
melanogaster and Z. vittiger) reared under standard, controlled conditions. Further, I sought to
determine whether estimates of the phenotypic plasticity of these traits remain constant over time in culture or diverge significantly under laboratory rearing conditions when acclimated at
different temperatures. Thus, I measured in both the F2 and F8-F10 generation of two species their
CTMAX, CTMIN, acute heat and chill survival after exposure to an extreme temperature,
desiccation resistance and starvation resistance and the plasticity of each trait in response to thermal acclimation at three temperatures (18, 23, 28 ˚C). I made two general predictions for the outcome of my study: I expected a decrease in both basal resistance and their plastic responses due to laboratory adaptation, possibly driven by small population sizes, as previously shown (Hoffmann et al., 2001; Sgró and Partridge, 2000). In addition, I expected there to be a decline in the plasticity of traits between the F2 and F10 generations if plasticity is costly to maintain.
Alternatively, if plasticity and basal stress resistance are traded-off directly, that basal stress resistance may decline in culture while plasticity could increase (or vice versa).
An additional objective of this work was to determine if population comparisons are subject to similar kinds of laboratory adaptation problems by identifying the nature and magnitude of trait variation and acclimation-induced phenotypic plasticity in four populations of D. melanogaster collected from different areas within South Africa. The F2 generation of each D. melanogaster
population was used to measure the same traits as for the species mentioned in the first objective. It is expected that there will be differences between environmental stress resistance of the
20
different D. melanogaster populations as there is known to be considerable variation between populations of the same species, especially if populations are sourced from climatically diverse habitats (David and Capy, 1988).
Aims:
1. To determine whether there is a large variation in environmental stress resistance traits and their plastic responses between newly established laboratory lines of D. melanogaster and Z. vittiger (i.e. at the 2nd generation) compared to lines that have been in laboratory culture for an extended period.
2. To determine the magnitude of trait variation and acclimation-induced phenotypic plasticity between four populations of D. melanogaster collected from geographically distinct areas within South Africa
21
Materials and Methods
Origin and maintenance of experimental flies
The Drosophila species used in this study were freshly collected from the field in Stellenbosch (STB), Brackenfell (BRAC), Durban (DB), the Cederberg (CD) and Polokwane (POL). For the species comparisons Drosophila melanogaster (STB) and Zaprionus vittiger (BRAC) were caught and compared and for the population comparisons Drosophila melanogaster from four different locations were sampled (STB, DB, CD and POL) (Figure 1).
Figure 1: Sampling locations for the Drosophila populations and species within South Africa
used in this study. Species (D. melanogaster and Z. vittiger) are represented by black
diamondss and populations of D. melanogaster (Pop1: Stellenbosch, Pop2: Polokwane, Pop3: Durban and Pop4: Citrusdal) by the crossed squares.
22
Flies were caught by placing traps consisting of buckets filled with mixed fruits (oranges, bananas, apples and lemons) in suitable locations and then collected on the rotting fruit with a hand net or a plastic bag. After collection, flies were placed individually in 325 ml plastic bottles containing Bloomington’s standard cornmeal medium
(http://flystocks.bio.indiana.edu/Fly_Work/media-recipes/bloomfood.htm). Mated wild- caught females were used to start eighteen isofemale lines for each of the two species and for each of the four populations (David et al., 2005). Flies were then placed in an incubator (MRC LE-509, Holon, Israel) pre-set at 23°C and <10% relative humidity on a 12h day/night cycle and tipped into fresh medium until the F2 generation when they were used for trait assays (Figure 2). Flies
were tipped into new food at regular intervals to avoid overcrowding the bottles and the density of flies in each bottle were kept constant. Within the first 24 hours after eclosion of the F2
generation the flies were put into new medium and three of the lines (chosen at random) were acclimated at 18°C, three at 23°C and three at 28°C for 48 hours. Acclimation temperatures were modified from constant temperate (19°C) and constant tropical (27°C) as used by Overgaard et
al., 2011. After the 48 hours acclimation, the flies were placed back into the incubator set at
23°C until they were between 5-7 days old (about 24 to 48 hours), which is standard practice in
Drosophila stress resistance studies (Kellermann et al., 2012a; Sgrò et al., 2010), and then
environmental stress resistance traits were scored. It was assumed that the effects induced in the 48 hour acclimation period will last several days (as shown by Loeschcke et al. (1997) in which pretreated flies remained more resistant for several days in Drosophila hydei and D.
melanogaster) since we were not particularly interested in any highly transient trait variation.
For each species, nine of the eighteen lines were used for trials when the lines reached the F2
generation and the remaining nine lines were kept in the laboratory until the tenth generation and then the same traits as for the F2 generation were assayed.
23
Environmental stress resistance
Temperature traits
CTMAX and CTMIN were determined by taking fifteen male and fifteen female 5-7 old flies (from
each line) from the F2 generation and placing them in eppendorf tubes (1.5ml). The eppendorf
tubes were placed in a foam “boat” which was then placed in a circulating programmable refrigeration bath (Huber CC-410wl, Huber, Offenburg, Germany) filled with water for CTMAX
and with ethanol for CTMIN (to prevent the bath liquid freezing at low temperatures). Fine-gauge
(36-SWG) Type T thermocouples connected to a hand held two channel digital thermometer (Fluke 54 series II, Fluke Cooperation, China) were placed inside one of the eppendorf tubes and
Figure 2: Experimental schematic diagram. For each species and population 18 isofemale lines were
established and reared at 23°C until the F2 generation after which lines were acclimated at 18°C, 23°C and
28°C (3 lines per acclimation) and returned to 23°C after 48 hours acclimation. After being acclimated for 48h, life history, temperature and survival traits were scored. The same procedure of acclimation followed by trait scoring took place at the F10 generation.
24
on the "boat" to measure the representative temperature inside the tubes (eppendorf temperature) and the temperature on the boat (surface temperature). For CTMAX the water bath was pre-set to
25°C (23°C in the eppendorf) and heated at 0.15°C/min (0.1°C/min ramping experience in the eppendorf). Flies were acclimated to the bath for 15min before ramping started. During ramping, flies were checked intermittently for coordinated movement and CTMAX were scored as the
temperature at which the fly lost all mobility (after all spasms has ceased and death ensues) (Lutterschmidt and Huthison, 1997). Flies that stopped all spontaneous movement were poked with a piece of fishing line to ensure that CTMAX was reached. For CTMIN the water bath was also
pre-set to 25°C (23°C in the eppendorf) and then cooled at 0.15°C/min (0.1°C/min in the boat). The ramping temperature of 0.1°C/min were chosen in order to replicate the rate used in studies determining lethal temperatures of Stock Center derived Drosophila species and thus allowing comparisons to be more readily made with some of the existing literature (Kellermann et al., 2012b, Overgaard et al., 2011). CTMIN were then scored as the temperature at which the fly lost
all mobility. As with CTMAX the flies were poked with a piece of fishing line to ensure that
CTMIN were reached.
Heat and Chill survival
Heat and cold survival were determined by placing fifteen male and fifteen female flies (5-7 day old) from each line individually into eppendorf tubes and placing them on the same foam boat setup described above in a pre-set programmable bath as for the thermal traits. Flies were placed at 38°C for 1h for the heat survival treatment and at 0°C for 2h for the cold survival treatment (as per Bechsgaard et al., 2013). After the treatments flies were placed into the incubator at 23°C for 24h after which survival was scored.
Survival traits
For desiccation resistance traits, 5-7 day old post-eclosion flies (15 males and 15 females) were placed individually into empty glass vials sealed with gauze. The vials were placed in an airtight desiccator (Duran 250mm, DIN 12491, Germany) and 80-90% of its volume filled with silica gel (Merck). The desiccator was placed in a dark incubator to suppress activity, at 23°C and <10% relative humidity (modified from Kellermann et al., 2012a). A hygrochron iButton (Maxim ibutton, Hygrochon Hi-Res (-20°C to +85°C) Acc 0.5°C, USA) was placed inside the desiccator
25
to confirm the temperature and relative humidity during the experiment. Survival was scored four to five times a day until the first fly died and was then scored hourly until all flies had died. For starvation resistance 15 male and 15 female 5-7 day old flies per acclimation were placed individually into glass vials containing 5ml 0.5% agar solution (Matzkin et al., 2009). The vials were sealed with moist cotton wool, to achieve relatively high level of humidity and placed in an incubator set at 23°C. Separate measurements have confirmed that this will typically achieve constant >95% humidity for several days. Mortality was scored at the same time each day until all flies had died.
As a control for desiccation and starvation resistance 15 male and 15 female five to seven-day old flies were placed in glass vials containing a standard cornmeal medium covered with gauze and placed in a desiccator at 23°C. Deaths were scored twice a week until the first death and then daily until all flies had died.
Statistical analyses
The effect of the acclimation regimes and the number of generations spent in the laboratory on environmental stress resistance of the Drosophila species and populations were determined in Rstudio version 1.0.136 (R core team, 2013) and Statistica version 13 (Statsoft, Tulsa,
Oklahoma, USA). All data were tested for normality by doing a Shapiro-Wilk normality test. All significance levels for tests were set at p<0.05.
Normality was tested by running a Kruskal-Wallis test and homogeneity of variances were determined by plotting the raw residuals over the predicted values. Variances were considered homogeneous if the residuals and the fitted values were uncorrelated. Depending on the outcome of these tests either a factorial ANOVA or generalized linear model (GLM) was run. If the data was non-parametric a generalized linear model factorial ANOVA was run with a poisson distribution and a log-link function. Overdispersion was checked and corrected for, if present. For the thermal limits of the species data, CTMAX or CTMIN was used as dependent variables in
the model and the independent variables were the generation (F2, F8-10), acclimation (18°C, 23°C,
26
used as dependent variables and the independent variables were population (Stellenbosch, Citrusdal, Durban or Polokwane), acclimation (18°C, 23°C, 28°C) and sex (male or female). For heat and cold survival of the species data, a generalized linear model (GLM) with a logit link function and a binomial distribution was run using the “MASS” package in R (Venables et al., 2002) to assess the main effects and interaction of generation, acclimation and sex on the proportion survival 24 hours after exposure to a potentially lethal temperature. The same was done for the populations of D. melanogaster but the effects of population source, acclimation and sex on the proportion of survival were tested. The assumptions of normally distributed data and independent errors were met for all data. A post hoc ANOVA of the GLM was run to examine significant effects and the main interactions.
For the desiccation and starvation survival experiments, Kaplan-Meier survival curves were drawn using the ‘survival’ package in R (Therneau and Grambsch, 2000)and showed the proportion survival over time for the different acclimation regimes (18°C, 23°C and 28°C). The Cox-proportional hazards model, also in the ‘survival’ package, was used to determine the dependency of survival time on the predictor variables of desiccation and starvation. For the species, the main effects and interactions of generation, treatment, acclimation and sex on survival over time were determined using the coxph function in the ‘survival’ package. For the populations of D. melanogaster, the main effects and interactions of origin of population,
treatments, acclimation and sex on survival time were also determined using the coxph function. The proportionality of hazards assumption for the cox regression was met in all analyses
(cox.zph function R). A post hoc ANOVA of the coxph model was run to examine significant effects and main interactions.
Results
Temperature treatments
Drosophila melanogaster
For CTMAX, the data were normally distributed (W=0.99, p>0.05) and the variances
homogeneous. There was a significant increase in basal resistance as well as plastic responses between the F2 and F10 generations (F=414.1, d.f.=1, p<0.0001). There was also a significant
27
difference between acclimation responses (F=56.1, d.f.=2, p<0.0001) with 18°C and 28°C having significantly higher resistance to thermal stressors in comparison to individuals reared at the optimum temperature of 23°C. CTMAX values for males and females also differed significantly
with females having a higher CTMAX (F=63.2, d.f.=1, p<0.0001). Both generation and
acclimation had a significant impact on CTMAX (F=4.2, d.f.=2, p<0.05) (Table 1; Figure 3).
For CTMIN, the data was not normally distributed (W=0.98, p<0.05) but the residuals were
homogeneous. The factorial GLM indicated a significant increase in CTMIN values between the
F2 and F10 generation of D. melanogaster, (Wald’s χ2=166.42, d.f.=1, p<0.0001). Significant
differences were also found between acclimations (Wald’s χ2= 56.1, d.f.=2, p<0.008) and the sexes (Wald’s χ2=9.64, d.f.=2, p<0.05). A significant interaction effect was found between
generation and acclimation (Wald’s χ2=48.6, d.f.=2, p<0.05) of CTMIN in D. melanogaster (Table
1; Figure 3).
Zaprionus vittiger
For CTMAX, data were non-normal (Shapiro-Wilks: W=0.95, p<0.05) and the variances were
heteroscedastic. There was a pronounced decrease in both basal resistance and plastic responses between the F2 and F10 generations (Wald’s χ2=190.3, d.f. =1, p<0.0001). There was a significant
effect of acclimation (Wald’s χ2=72.46, d.f. =2, p<0.0001) but not sex (Wald’s χ2=1.40, d.f.=1, p=0.24). Additionally, there was a significant interaction effect between generation and
acclimation (Wald’s χ2=20.68, d.f.=2, P<0.0001) and between acclimation and sex (Wald’s
χ2=17.59, d.f.=2, p<0.0002). The interaction between generation, acclimation and sex was
significant for CTMAX (Table 1; Figure 3). There was a significant decrease in CTMIN between the
F2 and F10 generations (Wald’s χ2=59.18, p<0.05). A significant effect of acclimation (Wald’s
χ2=50.58, p<0.05) and sex was detected (Wald’s χ2=32.29, p<0.05). There were significant
interaction effects between generation and acclimation (Wald’s χ2=39.82, p<0.05), generation
and sex (Wald’s χ2=33.60, p<0.05), acclimation and sex (Wald’s χ2=88.86, p<0.05) as well as
28
Figure 3: Box and whisker plot (mean ± SD) indicating the results of the temperature treatments
performed on the two species D. melanogaster (D. mel) and Z. vittiger (Z. vit). The figure shows the Critical thermal maximum (CTMAX) and Critical thermal minimum (CTMIN) results of D. melanogaster (first row) and Z. vittiger (second row) of the F2 and F10 generations for the three
acclimations of 18°C, 23°C and 28°C. Males are indicated by the black solid lines and unfilled circles; females are indicated by the grey stippled lines and filled squares.
29
Table 1: Temperature treatment results (CTMAX and CTMIN) from separate Factorial ANOVA and
Generalized linear models (GLM) for species (D. melanogaster and Z. vittiger) and populations of
D. melanogaster (significant effects indicated in bold).
Temperature treatments Effect F/ χ2 d.f p
D. melanogaster CTMAX Generation 141.1 1 <0.0001
Acclimation 56.1 2 <0.0001 Sex 63.2 1 <0.0001 Generation * Acclimation 4.2 2 <0.05 Generation * Sex 2.0 1 0.16 Acclimation * Sex 1.6 2 0.20 Generation*Acclimation*Sex 0.9 2 0.40
D. melanogaster CTMIN Generation 166.42 1 <0.0001
Acclimation 9.64 2 0.008 Sex 9.64 2 <0.05 Generation * Acclimation 20.68 2 <0.0001 Generation * Sex 0.16 1 0.69 Acclimation * Sex 45.28 2 <0.0001 Generation*Acclimation*Sex 5.36 2 0.069
Z. vittiger CTMAX Generation 190.34 1 <0.0001
Acclimation 72.46 2 <0.0001 Sex 1.40 1 0.24 Generation * Acclimation 104.42 2 <0.0001 Generation * Sex 0.81 1 0.37 Acclimation * Sex 17.59 2 <0.0002 Generation*Acclimation*Sex 5.24 2 0.072
Z. vittiger CTMIN Generation 59.18705 1 <0.0001
Acclimation 50.58265 2 <0.0001 Sex 32.28649 1 <0.0001 Generation * Acclimation 39.82445 2 <0.0001 Generation * Sex 33.59683 1 <0.0001 Acclimation * Sex 88.85645 2 <0.0001 Generation*Acclimation*Sex 39.49350 2 <0.0001
Populations CTMAX Population 114.30 3 <0.0001
Acclimation 12.65 2 <0.002
Sex 4.06 1 <0.05
30
Population*Sex 33.91 3 <0.0001
Acclimation*Sex 1.85 2 0.40
Population*Acclimation*Sex 51.05 6 <0.0001
Populations CTMIN Population 220.72 3 <0.0001
Acclimation 27.7046 2 <0.0001 Sex 0.0057 1 0.94 Population*Acclimation 18.2168 6 0.006 Population*Sex 144.4887 3 <0.0001 Acclimation*Sex 10.3262 2 0.006 Population*Acclimation*Sex 49.7351 6 <0.0001
Heat and Chill survival
Drosophila melanogaster
The GLM indicated acclimation to have a significant effect on heat survival (F=17.37, d.f.=2, p<0.0001). Significant interactions were found between the generations and acclimation
(F=12.43, d.f.=2, P<0.0001), acclimation and sex (F=4.08, d.f.=2, p<0.02) as well as generation, acclimation and sex (F=10.28, d.f.=2, p<0.0001). Flies acclimated to 23°C and 28°C had the highest percentage of individuals alive after 24h with the F2 males from the 23°C acclimation
having the highest percentage of individuals still alive (80%). For F10 flies, the highest survival
occurred in the 28°C acclimation group with 18°C having no survival and 23°C having some survival but only in the females. For F2 flies, highest survival was observed in the 23°C and 28°C
groups with the 18°C group having the lowest survival (Table 2; Figure 4).
For cold survival, acclimation also had a significant effect (F=7.22, d.f.=2, p<0.001). Significant interactions were found between generations and acclimation (F=33.28, d.f.=2, P<0.0001), generation and sex (F=11.63, d.f.=1, p<0.0009), acclimation and sex (F=8.51, d.f.=2, p<0.0004) as well as generation, acclimation and sex (F=5.28, d.f.=2, p<0.006) (Table 2). In F2 flies,
highest survival was observed in the 23°C and 28°C acclimation groups with the 18°C group having the least survival (as seen in the heat survival results). In the F10 generation however there
31
Zaprionus vittiger
Generation (F=21.54, d.f.=1, p<0.0001), acclimation (F=19.55, d.f.=2, p<0.0001) and sex (F=21.54, d.f.=1, p<0.0001) all significantly influenced heat survival in Z. vittiger. Significant interactions were also found between generation and acclimation (F=12.01, d.f.=2, p<0.0001), generation and sex (F=7.75, d.f.=1, p<0.006), acclimation and sex (F=19.76, d.f.=2, p<0.0001) and generation, acclimation and sex (F=9.85, d.f.=2, p<0.0001) (Table 2). F2 flies had a higher
survival in the 18°C and 23°C acclimation groups compared to F8-10 flies. In the 28°C
acclimation group F8-10 females had the highest survival and F8-10 males the lowest survival.
There were also significant differences between the sexes in the F8-10 generation with females
being more tolerant in the 18°C and 28°C acclimations and males being more tolerant for the 23°C acclimation (Figure 4).
For chill survival, acclimation (F=6.54, d.f.=2, p<0.002) and sex (F=6.46, d.f.=1, p<0.02) had a significant effect on recovery in Z.vittiger. In addition, there were significant interactions between all groups except for acclimation and sex (F=1.19, d.f.=2, p=0.31) (Table 2). In the F2
generation, 18°C had the highest survival followed by 28°C and lastly 23°C. In the F8-10
generation, however the highest survival was in the 28°C group followed by 23°C and lastly 18°C. In the F2 generation, males had higher survival than females in the 18°C and 28°C group;
and in the 23°C survival were similar. For the F8-10 generation males and females had similar
survival at 18°C acclimation, males had the highest survival at 23°C and females had the highest survival at 28°C (Figure 4).
32
Figure 4: Column graph indicating the mean survival (%) and standard error following heat (38°C for 1hour) and cold
survival (0°C for 2hours) treatments for Drosophila melanogaster and Zaprionus vittiger. Mean survival by Generation (F2=second generation and F10= tenth generation), acclimation (18°C, 23°C and 28°C) and sex (Males and Females) are
shown
33
Table 2: Heat and cold survival results from generalized linear models for species (D.
melanogaster and Z. vittiger) and populations of D. melanogaster (Significant effects indicated
in bold).
Heat survival Effect F d.f. p
D. melanogaster Generation 2.87 1 0.09 Acclimation 17.37 2 <0.0001 Sex 0.18 1 0.68 Generation*Acclimation 12.43 2 <0.0001 Generation*Sex 0.71 1 0.40 Acclimation*Sex 4.08 2 <0.02 Generation*Acclimation*Sex 10.28 2 <0.0001 Z. vittiger Generation 21.54 1 <0.0001 Acclimation 19.55 2 <0.0001 Sex 21.54 1 <0.0001 Generation*Acclimation 12.01 2 <0.0001 Generation*Sex 7.75 1 <0.006 Acclimation*Sex 19.76 2 <0.0001 Generation*Acclimation*Sex 9.85 2 <0.0001 Populations Population 7.00 3 <0.0002 Acclimation 4.041 2 <0.02 Sex 18.33 1 <0.0001 Population*Acclimation 16.26 6 <0.0001 Population*Sex 6.53 3 <0.0003 Acclimation*Sex 18.99 2 <0.0001 Population*Acclimation*Sex 8.22 6 <0.0001 Cold survival D. melanogaster Generation 0.00 1 1.00 Acclimation 7.22 2 <0.001 Sex 1.30 1 0.26 Generation*Acclimation 33.28 2 <0.0001 Generation*Sex 11.63 1 <0.0009 Acclimation*Sex 8.51 2 <0.0004 Generation*Acclimation*Sex 5.28 2 <0.006 Z. vittiger Generation 3.09 1 0.08 Acclimation 6.54 2 <0.002 Sex 6.46 1 <0.02 Generation*Acclimation 7.23 2 <0.001 Generation*Sex 11.05 1 <0.002 Acclimation*Sex 1.19 2 0.31
34 Generation*Acclimation*Sex 6.92 2 <0.002 Populations Population 18.43 3 <0.0001 Acclimation 6.87 2 <0.002 Sex 1.40 1 0.24 Population*Acclimation 15.56 6 <0.0001 Population*Sex 20.07 3 <0.0001 Acclimation*Sex 0.42 2 0.66 Population*Acclimation*Sex 13.55 6 <0.0001
Survival treatments
Drosophila melanogaster
The Cox-proportional hazards model showed no difference in survival between the generations (χ2=0.018, d.f.= 1, p>0.89). Although there were no differences between the generations overall,
differences were evident in the starvation treatments where the 18°C individuals in the F2 had a
LT50 of 160.45 versus F8-10 at 67.17, 23°C individuals at F2 had a LT50 of 230.08 versus F8-10 at
53.85 and 28°C F2 survived 156.84 versus F8-10 at 51.18 and thus there was a significant decrease
between the generations for starvation resistance (χ2=137.11, d.f.=1, p<0.0001). The model indicated significant differences between the treatments (χ2=830.64, d.f.=2, p<0.0001). The desiccation treatment survived the shortest amount of time (median=14.5 h), followed by
starvation (median=105.5 h) and then the control (median=519.5) (χ2=830.64, d.f.=2, p<0.0001). There was no difference between the acclimation groups (χ2=1.4052, d.f.=2, p=0.50), but there were differences between the sexes (χ2=28.38, d.f.=28.381 p<0.0001) for the starvation
treatment. In both generations females survived significantly longer than males except in the 28°C acclimation of the F2 generation where males performed better than females (Figure 5).
Generation, treatment, acclimation and sex had an effect on the survival rate of D. melanogaster (Table 3; Figure 6a; Figure 7).
35
Zaprionus vittiger
The Cox-proportional hazards model indicated a significant difference between the generations (χ2=56.21, d.f.=1, p<0.0001), treatments (χ2=586.11, d.f.=1 p<0.0001) and acclimations
(χ2=17.56, d.f.=2, p<0.0002) in Z. vittiger. Significant interaction effects were detected between
generation and acclimation (χ2=43.15, d.f.=2, p<0.0001) as well as generation, treatment, and
acclimation (χ2=6.09, d.f.=2, p<0.05) (Table 3). The basal resistance for the desiccation,
starvation and the control of the F2 generation was significantly higher than that of the F10
generation in Z. vittiger. The acclimation responses of the F2 generation were also higher than
that of the F8-10 generation. Zaprionus vittiger flies that underwent desiccation trials had an LT50
basal survival at the F2 generation of 37 hours versus 16 hours at the F8-10 generation, those
undergoing starvation trials had an LT50 basal survival value of 140 hours at the F2 versus 73
hours at F8-10. The control flies had a basal LT50 survival of 452 hours at F2 versus 431 hours at
F8-10. Plastic responses were lower than the basal resistance in the F2 generation for desiccation
and starvation resistance. However, for the control in the F2 generation the 18°C acclimation and
the basal (23°C) were similar and the 28°C acclimation flies survived the longest. For the F10
generation the 18°C acclimation had the highest survival rate for both desiccation and starvation resistance. The control group of the F10 also has the highest survival at the 28°C acclimation
36
Table 3: Survival treatment results from a Cox-proportional hazards model, indicating χ2, degrees
of freedom and p-value for D. melanogaster, Z. vittiger and the comparison between the four D.
melanogaster populations (Significant effects indicated in bold).
Species Effect χ2 d.f. p D. melanogaster Generation 0.0181 1 0.89 Treatment 830.64 2 <0.0001 Acclimation 1.4052 2 0.4952 Sex 28.381 1 <0.0001 Generation*Treatment 121.898 2 <0.0001 Generation*Acclimation 20.836 2 <0.0001 Treatment*Acclimation 27.8008 4 <0.0001 Generation*Sex 0.7468 1 0.39 Treatment*Sex 7.2803 2 <0.05 Acclimation*Sex 9.2319 2 <0.01 Treatment*Acclimation*Sex 24.76 4 <0.0001 Generation*Treatment*Acclimation 33.91 4 <0.0001 Generation*Treatment*Sex 27.7158 2 <0.0001 Generation*Acclimation*Sex 0.41 2 0.814 Generation*Treatment*Acclimation*Sex 10.50 4 <0.05 Z. vittiger Generation 56.21 1 <0.0001 Treatment 586.11 2 <0.0001 Acclimation 17.5567 2 <0.0002 Sex 2.9741 1 0.085 Generation*Treatment 0.7727 1 0.38 Generation*Acclimation 43.1480 2 <0.0001 Generation*Sex 1.4025 1 0.24 Treatment*Acclimation 4.1159 2 0.13 Treatment*Sex 0.2645 1 0.61 Acclimation*Sex 0.3790 2 0.83 Treatment*Acclimation*Sex 1.0746 2 0.58 Generation*Treatment*Acclimation 6.0932 2 <0.05 Generation*Treatment*Sex 0.0111 1 0.92 Generation*Acclimation*Sex 0.7922 2 0.67 Generation*Treatment*Acclimation*Sex 3.1029 2 0.21 Populations Population 23.02 4 <0.0002 Treatment 1990.76 2 <0.0001 Acclimation 3.44 2 0.18 Sex 25.87 1 <0.0001 Population*Treatment 305.46 6 <0.0001 Population*Acclimation 23.25 6 <0.001
37 Treatment*Acclimation 31.92 4 <0.0001 Population*Sex 5.62 3 0.13 Treatment*Sex 9.20 2 <0.05 Acclimation*Sex 8.45 2 <0.05 Population*Treatment*Acclimation 65.38 12 <0.0001 Population*Treatment*Sex 29.37 6 <0.0001 Population*Acclimation*Sex 14.00 6 <0.05 Treatment*Acclimation*Sex 5.66 4 0.23 Population*Treatment*Acclimation*Sex 22.24 12 <0.05
Figure 5: Column graph indicating the mean starvation resistance survival (hours) and
its standard error in D. melanogaster at the three acclimations (18°C; 23°C; 28°C), for the two generations (F2; F10) indicating male and female responses
38
Figure 6: Kaplan-Meier Survival curves indicating a) the survival curves with standard error (grey area) for D.
melanogaster depicting both generations (F2 and F10), all treatments (C=Control, D=Desiccation, S=Starvation) and the
three acclimations (18°C, 23°C, 28°C), b) the survival curve with standard error for the Desiccation treatment of D.
melanogaster depicting both generations (F2 and F10 and acclimations (18°C, 23°C, 28°C)), c) the survival curve with
standard error for Z. vittiger depicting both generations (F2 and F10), all treatments (C=Control, D=Desiccation, S=Starvation) and the three acclimations (18°C, 23°C and 28°C) and d) the survival curve with standard error for the Desiccation treatment of Z. vittiger depicting both generations (F2 and F10 and acclimations (18°C, 23°C and 28°C)).
39
Figure 7: Column graph showing 50% survival (in hours) and 95% confidence limits for
desiccation and starvation resistance as well as a control at 18°C, 23°C and 28°C for the second and tenth generation (F2 and F10) in D. melanogaster and Z. vittiger.
40
Populations of D. melanogaster
Temperature treatments
For CTMAX, a Shapiro-Wilk normality test was performed, the data were normally distributed
(W=0.028, p>0.05) and variances were homogeneous. There were significant differences between the populations (F=114.30, d.f.=3, p<0.0001), acclimation (F=12.65, d.f.=2, p<0.002) and sex (F=4.06, p<0.05). There were also significant interactions between population and acclimation (F=70.18, d.f.=6, p<0.0001), population and sex (F=33.91, d.f.=3, p<0.0001) and population, acclimation, and sex (F=51.05, d.f.=6, p<0.0001) (Table 1). For CTMAX, the basal
resistance (23°C) of the Stellenbosch population was the highest, followed by Citrusdal, Durban and Polokwane. There were also diverse plastic responses (acclimations 18°C and 28°C)
between the populations. The 18°C acclimation in the Citrusdal population was similar to the basal resistance (23°C) but for the 28°C acclimation group females had a very low resistance (mean CTMAX of 35°C) whilst the males had a higher CTMAX of 38°C. For the Polokwane
population, the plastic responses were very similar to the basal plastic response with a 1°C increase or decrease. Stellenbosch D. melanogaster had higher resistance in the 18°C and 28°C groups. Lastly, the Durban population looked similar to the Citrusdal population with the 18°C group being similar to the 23°C group and the 28°C group being very different, but here the females were more resistant with females having a CTMAX of 39°C and males having a CTMAX of
36°C (Figure 8a).
For CTMIN, a generalized linear model was run after data were found to not be normally
distributed (W=0.99, p<0.05) and the variances heterogeneous. Populations were significantly different from each other (χ2=220.72, d.f.=3, p<0.0001) as well as the different acclimation
groups (χ2=27.70, d.f.=2, p<0.0001). There were also significant interactions between population and acclimation (F=18.2168, d.f.=6, p=0.006), population and sex (F=144.4887, d.f.=3,
p<0.0001), acclimation and sex (F=10.3262, d.f.=2, p=0.006) and population, acclimation and sex (F=49.7351, d.f.=6, p<0.0001). Thus, population, acclimation and sex all had significant effects on CTMIN (Table 1). Durban had the lowest CTMIN and Stellenbosch the highest. Citrusdal
and Polokwane averaged between 3°C and 7°C. Resistance of the plastic responses (18°C and 28°C) was higher than the basal resistance in the Citrusdal D. melanogaster population. In addition, resistance of the female flies was lower than the males in the Citrusdal population
41
(higher CTMIN). Durban had the opposite response to Citrusdal with males having lower
resistance (higher CTMIN) (Figure 8). In the Durban population plastic responses also had higher
resistance than the 23°C group, except for the 23°C females which had a lower resistance than the 23°C group. The Stellenbosch D. melanogaster had a lower basal resistance in both sexes except for 28°C acclimated males which had a higher basal resistance than the 23°C group. Lastly for the Polokwane population, the males had higher resistance in the acclimation groups 18°C and 28°C than in the 23°C group whilst the females had higher resistance in the 23°C group than in the plastic groups (Figure 8; Table 1).
42
Figure 8: Box and whisker plots indicating mean and standard deviation of a) CTMAX and b) CTMIN of the four D. melanogaster populations (Citrusdal, Polokwane, Stellenbosch and Durban),
indicating acclimations (18°C, 23°C and 28°C) and sex (Male and Female). The figure shows the Critical thermal maximum (CTMAX) (first column) and Critical thermal minimum (CTMIN) (second
column) results of the four D. melanogaster of the F2 generation at the three acclimations of 18°C,
23°C and 28°C. Males are indicated by the black solid lines and unfilled circles; females are indicated by the grey stippled lines and filled squares.
